OCTOBER 1966
CONDENSATION OF DIKETONES WITH MONOKETONES
3159
(Y
Monoketones as Acceptors in the Oxyphosphorane Carbon-Carbon Condensation. Condensation of a Diketones with Monoketones and with Perfluoro Ketones. P3* Nuclear Magnetic Resonance Spectra”2 FAUSTO RAMIREZ,~ A. V. PATWARDHAN, AND C. P . SMITH Department of Chemistry, State University of New York at Stony Brook, Stony Brook, New York Received May 27, 1966 The 2,2,2-trimethoxy-l,3,2-dioxaphospholenemade from biacetyl and trimethyl phosphite reacted with cyclohexanone, cyclobutanone, acetophenone, p-nitroacetophenone, and a,a,a-trifluoroacetophenone, at about 100120’. The reaction with hexafluoroacetone was rapid even below 10’. These oxyphosphorane condensations proceeded by addition of a carbon atom of the phospholene to the carbonyl carbon of the monoketone, with formation of a 2,2,2-trimethoxy-1,3,Z-dioxaphospholane. When the monoketone had different substituenb about the carbonyl, two diastereomeric phospholanes were formed. The ratio of diastereomers could be altered by using an excess of the ketone or an excess of the phospholene as the reaction medium. The formation of some phospholanes was reversible at elevated temperatures. The pyrolysis of the phospholene gave the trialkyl phosphite, the a diketone, trialkyl phosphate, and other products. The following monoketones failed to react with the dioxaphospholene: cyclopentsnone, acetone, butanone, 3-hexanone, methyl cyclopropyl ketone, and benzophenone. The PSInmr shifts of the phospholanes were measured at 40.5 Mcps.
A previous paper in this series described the reaction of the 2,2,2-trimethoxy-l13,2-dioxaphospholene (I) made from biacetyl and trimethyl phosphite, with propionaldehyde and with other aliphatic monoaldehydes.* These reactions occurred at 20°, and gave one diastereomer of the 2,2,2-trimethoxy-1,3,2-dioxaphospholane (11).
I, aPp1=+ 48.9 ppm
11,
+ 51.3 ppm
Configuration 11, with the two alkyl groups in trans relationship, was based, in part, on the positions of the H1 nmr signals due to the acetyl and the methyl groups. Considerable information was available on the effect of neighboring groups on the positions of these signals in other related oxyphosphoranes.2,5 Configuration I1 was assigned also by analogy with the results of the reaction of the phospholene I with benzaldehyde.2 In this case, two diastereomers, IIIa and IIIb, were formed in the proportion of 9O:lO. The H’ nmr signal due to the acetyl group was at a higher magnetic field in the major isomer ( T 8.40) than in the minor isomer ( T 7.76). This suggested that the acetyl group was cis to the phenyl ring in the
major isomer, IIIa, and trans to the phenyl ring in the minor isomer, IIIb. The pentaoxyphosphoranes I1 and I11 gave positive P31 nmr shifts relative to H 3 P 0 4 . I n the benzaldehyde case, the minor isomer could not be detected in the P31 nmr spectrum. However, in a number of other cases already reported from this l a b o r a t ~ r y , ~ the Pal nmr shifts of the two diastereomeric phospholanes analogous to IIIa and IIIb were easily distinguished at a frequency of 40.5 Mcps. For example, methyl pyruvate reacted with trimethyl phosphite and gave two diastereomeric phospholanes in a 40 :60 proportion.6 The minor isomer, which had two methyl groups in cis relationship gave a decet centered at +52.2 ppm; the major isomer (racemic) had its decet displaced 1.8 ppm to lower field.5 The present paper extends our studies of the oxyphosphorane carbon-carbon condensation to the case of aliphatic, alicyclic, and aromatic monoketones. Various types of ketones with and without polar substituents were studied in order to ascertain more fully the stereoelectronic factors which influence the condensation, and to delimit its synthetic scope. Results Reaction of Alicyclic Monoketones with Dioxaphospho1enes.-Cyclohexanone reacted with the dimethyldioxaphospholene I under the conditions given in Table I. The P31 and the H’ nmr shifts of the resulting dioxaphospholane,IV, are listed in Table 11. Note,
11
CH3 CHs
IIIa, +51.5 ppm
IIIb
(1) Organic Compounds with Pentavalent Phosphorus. Part XXIII. Supported by Public Health Service Reaearch Grant No. CA-04769-07from the National Cancer Institute and by the National Science Foundation, GP-3341. (2) Part XXII: F. Ramiree, A. V. Patwardhan, and C. P. Smith, J . Org. Chem., 81, 474 !1966). (3) National Science Foundation Senior Postdoctoral Fellow, 1966-1966. (4) F. Ramiree, A. V. Patwardhan, N. Ramanathan. N. B. Desai. C. V. Greco, and S. R. Heller, J . A m . C&m. Soc., 87, 543 (1965). (6) F. Ramirer, Pure AppE. Chem., 8 , 337 (1964).
I
0 CH3 IV, 1-54.3 ppm
in particular, the large positive value of the P nmr shift. These values and the infrared data given in the Experimental Section establish structure IV, as discussed in previous papers of this series.2,44 (6) F . Ramiree, N. B. Desai, and N. Ramanathan, Tetrahedron Lettem,
No. 5 , 323 (1963).
RAMIREZ, PATWARDHAN, AND SMITH
3160 REACTION OF
THE
TABLEI TRIMETHYL BIACETYLPHOSPHITE 1: 1 ADDUCT' WITH
Expt
1
2 3 4 5 6 7 8 9 10 11
12 13 14 15 16 17
Monoketone
Cyclohexanone Cyclohexanone Cyclohexanone Cyclohexanone Cyclohexanone Cyclohexanone Cyclohexanone Cyclopentanone Cyclobutanone Cyclobutanone Cyclobutanone Cyclobutanone Acetone Butanone 3-Hexanone Methyl cyclopropyl ketone Hexafluoroacetone
MONOKETONES
Mole ratiob
5 3 3 1 3 3 5 3 3 3 3 4
5 3 3 3 f
Temp, oC0
95 100 130 155' 158 158 20 135 100 100 100 20 20 80 140 100 -7Oto +20 125 100 100 100 95
Dioxaphsph*d lane,
%
Time
7days 40hr 1 hr 3hr 3hr 6 hr 6days 6hr 40hr 5days 3hr 6days 3months 48hr 40hr
78 42 25 40 42 45 0 0 40 53 15 0 0 0 0
40hr 3 hr
0 90
+
+
TABLE I1
Psl
AND
H' NMRSHIFTS'
OF
2,2,ZTRIMETHOXY-1,3,2.DIOXA-
CONDENSATION OF THE TRIMETHYL BIACETYLPHOSPHITE 1: 1 ADDUCT"(I) WITH MONOKETONES
PHOSPHOLANES~ FROM THE
Acetophenone 3 72hr 500 16hr Ca. 5 Acetophenone 1 p-Nitroacetophenone 2 72 hr h p-Nitroacetophenone 0 . 5 64 hri 65j p-Nitroacetophenone 0 . 3 24 hr 3ot a-Trifluoroacetophe 1 95 20hr 85' none 24 a-Trifluoroacetophe none 1 20 4days 30 25 Benzophenone 1 140 24hr 0 a 2,2,2 - Trimethoxy - 4,5 - dimethyl - 1,3,2 - dioxaphospholene. Monoketone-adduct (1: 1). c Internal temperature, f 5 ' . d Distilled or crystallized. 6 Initial reflux temperature; final reflux temperature, ca. 170'. f An excess of gaseous ketone was condensed into a solution of adduct in hexane a t -70". 0 Only one isomer with the acetyl trans to the phenyl could be detected. * Product not isolated. H1 nmr showed mostly the phospholane with acetyl trans to p-nitrophenyl. Minor amounts of cis isomer and of (CHs0)aPO were observed. Little change after 40 hr. i Crystalline mixture of trans and cis isomers (IXa IXb) in 20: 80 proportion. Original crude mixture of IXa IXb in 35:65 proportion (90% yield). Pure cis isomer isolated in 41% yield. Pure cis isomer IXb. Liquid mixture of trans and cis isomers Xa Xb in 82:18 proportions. All isomer proportions from integration of HI nmr signals. 18 19 20 21 22 23
VOL.31
+-
Monoketone acceptor Cyclohexanone Cyclobutanone Hexafluoroacetone Acetophenone pNitroacetophenone a-Tri5uoroacetophenone
bP"
No. IV VI VI1 VI11 IXa
+54.3 4-53.1 +52.6 4-54.2 4-54.0 IXb 4-52.4 Xa 52.5 {Xb +49.6
+
z~~~ T
M
~THE' ~
7.83 8.85 None0 7.82 8.80 Noneh 7.72( 8 . 5 1 j None 7.7ok 9.15' 8.55 7.60'" 9.12' 8.53 8.37" 8.50 8.30 7.BOb 9.00' None 8.42" 8.25' None
TM~O
Jmf
6.44 6.45 6.38 8.40 6.32 6.34 6.35
12.5 12.8 13.0 12.8 12.8 12.5 12.5
... ...
P shifts in parts per million from 85% HaPOr = 0; measured a t 40.5 Mcps in CC4 or CHaC12. H shifts in parts per million from TMS = 10 ( 7 values); measured a t 60 Mcps in CClr or CDCla. 6 Diastereomers a t carbon are possible with certain monoketones. e 2,2,2-Trimethoxy-4,5-dimethyl-l,3,2-dioxaphospholene. Acetyl and methyl groups derived from biacetyl. 6 Methyl group derived from monoketone. f Coupling in cycles per second, between methoxy protons and phosphorus. 0 The ten methylene protons gave a broad signal a t ca. 7 8.4. The six methylene protons gave a signal a t ca. 7 7.8. i Quartet, Jm = 0.8 cps. j Quartet, J m = 2 cps; see Figure 1. k Acetyl trans to phenyl. 6 Methyl cis to phenyl, hence signal a t higher magnetic field. Acetyl trans to p-nitrophenyl. Acetyl cis to p-nitrophenyl, hence signal a t higher magnetic field. Acetyl cis to phenyl. P Methyl is broad owing to cis-trifluoromethyl.
CH3 CHs
ha
o
+ I -
Vb, t52.1 ppm
Va, +64.3 ppm
The origin of the trimethyl phosphate is being investigated further. The major product of the decomposition of the phospholane IV after 6 hr at 165" was cyclohexanone; i.e., the formation of IV was reversible under these conditions. Cyclopentanone failed to condense with the phospholene I; cf. expt 8 of Table I. Cyclobutanone, however, gave the corresponding phospholane VI, as shown in expt 9 and 10. The nmr shifts are given in Table 11.
Experiments 1 to 7 of Table I disclosed that the temperature was an important factor in the condensation. Significant reaction was noted only above 90"; however, there was a rather narrow limit to the useful temperature range, because both the phospholene I and the phospholane IV began to decompose, slowly, at about 130". The best conditions for making the latter, IV, were those of expt 1 and 5. Several products were formed when the phospholene I was kept for 72 hr at 125": (a) trimethyl phosphate, 0 CHa (b) trimethyl phosphite, and (c) the meso and racemic VI, + 53.1 ppm forms of the 2,2,2-trimethoxy-4,5-dimethyl-4,5-diacetyl1,3,2-dioxaphospholane (Va and Vb). The latter had Reaction of Aliphatic Monoketones with Dioxabeen previously made from the phospholene I and phospho1enes.-Experiments 13 to 16 disclosed that biacetyl.' Therefore, part of the pyrolysis of the most of these ketones did not engage in the oxyphos1: 1 adduct I proceeded as shown. phorane condensation. On the other hand, hexa(7) (a) F. Ramiree, N. Ramanathan, and N. B. Deaai, J . Ana. Chmn. Soc.. fluoroacetone proved to be the most reactive of all 84, 1317 (1962);(b) F. Ramirez and N. B. Desai, ibid., 86, 3252 (1963); the monoketones studied; this is shown in expt 17. (0) F. Rsmirez, N. Rsmanathsn, and N. B. Desai, ibid., 85, 3465 (1963).
*
OCTOBER1966
CONDENSATION OF a DIKETONES WITH MONOKETONES
The ninr shifts of the phospholane VI1 are given in Table 11. The H1 nmr spectrum of VI1 is reproduced in Figure 1, to show the long-range coupling of the H' and the F'9 nuclei. Long-range HI-FIg couplings have been observed previously in other systems.* Reaction of Aromatic Monoketones with Dioxaphospho1enes.-Acetophenone reacted with the dimethyldioxaphospholene I as shown by expt 18 of Table I. Only one of the two possible diastereomeric phospholanes could be isolated in the case. This isomer is thought to have the acetyl group in trans relationship to the phenyl ring as in formula VIII.
fi
VII, +52.6 ppm
VIII, +54.2 pprn
The basis for this configurational assignment is the position of the signal due to the protons of one of the methyl groups in the molecule. As shown in Table 11, this signal is at a relatively high magnetic field (7 9.15); this is understandable if the methyl is cis to the phenyl ring.2~4J p-Nitroacetophenone was an excellent acceptor in the oxyphosphorane condensation. Preliminary experiments were carried out with an excess of the ketone, as shown in Table I, expt 20. Under these conditions, the major product was the phospholane IXa with the acetyl group trans to the p-nitrophenyl ring. This configurational assignment foliowed also from the nmr data shown in Table 11. Note the relatively high position ( r 9.12) of the methyl group cis to the p-nitrophenyl ring. The crude reaction mixture contained small amounts of the isomer IXb with cis-acetyl p-nitrophenyl groups. In addition, some trimethyl phosphate had been produced.
3161
Figure 1.
Note the high position, r 8.37, of the H1 nmr signal due to the acetyl group cis to a p-nitrophenyl ring. The different isomer ratio obtained under different experimental conditions seemed to indicate that the phospholanes IXa and IXb were formed reversibly at elevated temperatures. Indeed, the pure cis isomer (IXb) lost p-nitroacetophenone and regenerated the phospholene I at 120°, although rather slowly. The pyrolysis gave also small amounts of the trans isomer IXa. I n addition, trimethyl phosphate and a fifth substance were produced; the latter had a H1 nmr signal at r 7.82, but it was not further characterized at this time. The pyrolysis of these, and of other phospholanes, is under investigation. a,a,a-Trifluoroacetophenone was an excellent acceptor in the oxyphosphorane condensation, as shown by expt 23 of Table I. This monoketone was a rather volatile liquid and could be used as the reaction medium and then recovered. Under these conditions the diastereomers Xa and Xb were formed in the proportion of 82: 18 and in 90% yield or better. Table I1 lists the corresponding nmr shifts. The minor isomer Xb exhibited a small HI-FI9 long-range coupling which is consistent with the cis relationship of the CH3 and the CF3 groups. CF3
Xb, i 4 9 . 6 ppm
Xa, +52.5 ppm IXa, +54.0ppm
IXb, +52.4 ppm
Since p-nitrophenylacetophenone is a solid, and since it was difficult to remove from the product, subsequent experiments were carried out with an excess of the liquid phospholene I, as shown in expt 21 and 22 of Table I. Under these conditions, the major product was the phospholane I X b with the acetyE group cis to the p-nitrophenyl group. The isomers IXa IXb were formed in the proportion of 35:65 and in 90% yield. One crystallization from hexane gave a mixture of IXa IXb in a 20:80 proportion and in 65% yield. One crystallization from benzene-hexane gave the major isomer, IXb, in pure form (40% yield).
+
+
(8) (a) A. H. Lewin, J . Am. Chem. Soc., 86, 2303 (1964); (b) A. D. Croes and P. W. Landis, ibid., 86, 4005 (1964); (c) A. D. Cross, ibid., 86, 4011 (1964).
Discussion This investigation showed that monoketones are, in general, much less reactive than monoaldehydes toward dioxaphospholenes. However, electron-withdrawing substituents greatly enhanced the reactivity of the ketones; thus, hexafluoroacetoneg and trifluoroacetophenone were among the most reactive acceptors so far encountered in the oxyphosphorane condensation. The thermal decomposition of the phospholenes and of the phosphoianes imposed certain limitations on the scope of the oxyphosphorane condensations with the (9) Fluorine activation of carbon-carbon double bond toward nucleophiles has been discussed by W. T. Miller, Jr., J. H. Fried, and H. Goldwhite, ibid., 84, 3091 (1960), and previous papers: A. L. Henne and M. Nager. ibid., 74, 650 (1952); D. C. England, R. V. Lindsey, Jr.. and L. R. Melby, ibid., 80, 6442 (1958).
VOL.31
RAMIREZ, PATWARDHAN, AND SMITH
3162
less reactive ketones. These decompositions became significant above 130” and were quite destructive at about 170”. The oxyphosphorane condensation can be pictured as a concerted process in which a P-0 bond breaks as the new C-C bond forms.2Jo This is shown in transition states X I and XII. Little or no charge separation is required, since the new phospholane ring can be formed as the existing phospholene ring is being broken. These reactions occurred best in solvents of low polarity. The available data2g4-7J0*11suggest that the interplay of steric12 and electronic factors associated with the four groups, R’, R”, R”’, and R””, in the transition states X I and XI1 will determine (1) the degree of reactivity of the carbonyl compound, (2) the direction of the condensation when the nature of the groups R permits the formation of structural isomers, and (3) the proportion of the two diastereomers, trans and cis of a given structure. The relative reactivity of the cyclic ketones seemed to conform with the general behavior of these compounds toward nucleophile^.^^ The most surprising effect is, perhaps, the greater reactivity of acetophenone over the simple aliphatic monoketones like 3-hexanone. In this connection, it should be noted that 1-phenylpropanedione was a more reactive acceptor than benzil in the oxyphosphorane condensation; yet, the benzoyl carbon rather than the acetyl carbon was involved in the condensation. Isomers XIIIa (cis) and XIIIb (trans) were formed in the proportion 70: 30? These results were attributed partly
XIIIa, $54.7 ppm
tion, i.e., to the nature of the preferred structural isomer. The predominance of the cis isomer XIIIa over the trans isomer XIIIb was in agreement with the hypothesis of a significant dipoledipole interaction involving the two carbonyl groups in the transition sfate.l4 The isomer ratio, again, suggested a relatively small repulsion between a methyl and a phenyl in the transition state.l6 When the role of acceptor in these condensations is taken by a monocarbonyl compound, the main, if not the only factor involved is likely to be the relative steric repulsions12J6 of the four groups, R’, R”, R”’, and R””, in X I and XII. The lack of reactivity of the purely aliphatic ketones may simply reflect strong repulsions among the alkyl groups in the transition state. The greater reactivity of acetophenone is understandable if the methyl-phenyl repulsion is smaller than the methyl-alkyl repu1si0n.l~ As to the stereochemical results observed in this work, the evidence tentatively suggests that the kinetically favored diastereomer has the acetyl group trans to the phenyl ring. This is consistent with the relatively small cis methyl-phenyl repulsion. In the p-nitroacetophenone case, the thermodynamically favored diastereomer appeared to have the acetyl group cis to the p-nitrophenyl ring. This could be due, in part, to a dipole-dipole attraction between the carbonyl and the nitro group. However, a second factor could be related to the shape of the group R”” = CaH4. NOz-p and to the steric requirements of a trigonal bipyramid. The latter is assumed t o represent the geometry of the dioxaphospholane in solution. X-Ray analysis of a dioxaphospholene disclosed that the molecule in the crystal is a trigonal bipyramid with the five-membered ring in an apicalequatorial plane.l6 Models show that there is considerable crowding involving group R”” and the other groups attached to the phosphorus in a configuration like X I (transacetyl/R””). It seems quite possible that the crowding due to group R”” in the trans configuration is
XIIIb, +52.4 ppm
to the stronger attraction between two acetyl groups when compared with the attraction between an acetyl and a benzoyl group in the transition state of the condensation. This should lead to attack at the benzoyl rather than at the acetyl group. Moreover, it was pointed out that steric factors, in the form of methyl-phenyl and methyl-methyl repulsions, might be contributing also to the direction of the condensaR‘
(10) F. Ramirez, A. V. Patwardhan, and C. P. Smith,
J. Ow. Chem., S O ,
2575 (1965). (11) (a) F. Ramirez, H. J. Kugler, and C. P. Smith, Tetmhedron L e t t c ~ s , NO.4,261 (1965); (b) F. Ramirez and N . Ramanathan, J . O w . Cham., S 6 , 3041 (1961); (c) F. Ramirez, A. V. Patwardhan, N. B. Desai, N. Rammathan, and C. V. Greco, J. A m . Chem. SOC.,86, 3056 (1963). (12) For a recent discussion of the differences between “steric requirements” of groups, and group sizes measured b y van der Waals radii, see (d H. C. Brown and R. L. Khimisch, ibid., 88, 1425 (1966); (b) H. C. Brown, J. Chem. Educ., 86, 424 (1959). (13) (a) H. C. Brown, J. H. Brewster, and H. Shechter, J . A m . Chem. Soc., 76, 467 (1954); (b) M.8. Newman, “Steric Effects in Organic Chemistry,” M. 8. Newman, Ed., John Wiley and Sons, Inc., New York, N. Y., 1956, p 237.
XII, C~$-R’CO/R”” , CZ’S-R’‘/R)‘’ (14) All previous oxyphosphoranes condensations were carried out at 20’ or below. It was shown that, under those conditions, the condensations were irreversible. There was no detectable equilibration among diaatereorners, once they were formed. See ref 2 , 4 , 6 , 7, 10, and 11. (16) The van der Waals radius (“size”) of the methyl group is said to be 2.0 A, while the half-thickness of the aromatic ring has been estimated as 1.85 A; cf. L. Pauling, “The Nature of the Chemical Bond,” 3rd ed, Cornel1 University Press, Ithaca, N . Y., 1960. However, see footnote 9 for difficulties in estimating steric requirements of groups from their “sizes.” (16) W. C. Hamilton, 8. J. LaPlaca, and F. Ramirez, J . A m . Chem. SOC. 67, 127 (1965).
OCTOBER 1966
CONDENSATION OF CY DIKETONES WITH MONOKETONES
more significant in the product IX than in the corresponding transition state because the former might be closer than the latter to being a, pure trigonal bipyramid. It is becoming apparent that the formation of carboncarbon bonds via oxyphosphoranes is a rather general reaction. The dioxaphospholanes can be hydrolyzed to cyclic and open-chain phosphate esters and to phosphorus-free polyhydroxycarbonyl compounds. This paper constitutes a significant extension of the synthetic scope of this new reaction.
Experimental Section Psl nmr shifts are given in parts per million vs. 85% HsPOd, taken a t 40.5 Mcpse4 H1 nmr shifts are given in parts per million from tetramethylsilane = 10 ( r values), taken in a Varian A-60 i n ~ t r u m e n t . ~Analyses were performed by Schwarzkopf Microanalytical Laboratory, Woodside, N. Y. Reaction of the Trimethyl Biacetylphosphite 1:1 Adduct (I)4 with Cyclohexanone. Preparation of 2,2,2-Trimethoxy-4acetyl-4-methyl-5 - spiropentamethylene - 1,3,2-dioxaphospholane (IV).-A mixture of 1 :1 adduct I4(10.5 g, 50 mmoles) and freshly distilled cyclohexanone (25 g, 250 mmoles) was kept at 95-loo', under Nz. Analysis of the mixture by H1 nmr spectrometry showed about 60% reaction after 4 days and ca. 75% reaction after 7 days. Some trimethylphosphate began to appear a t the longer reaction time. Unreacted I and the excess cyclohexanone were removed a t 45-50' (20 mm; bath a t 100'). The cyclohexanonetrimethyl biacetylphosphite adduct IV (12.1 g, 78% yield) was collected a t 90-92" (0.05 mm; bath a t 135140') using a 6-in. Vigreux column: nZ6D1.4641. Anal. Calcd for C13H2606P:C, 50.6; H, 8.1; P, 10.1; mol wt, 308. Found: C, 50.4; H , 8.0; P, 9.9; mol wt, 261 (cryoscopic in benzene). The Psl (in cc14) and the H1 (in cc14) nmr shifts are listed in Table 11. The infrared spectrum (CCla) had a C=O band a t 5.84, and the strong and broad absorption at 9.3 fi due to the oxyphosphorane POCHs groups. Effect of Temperature, Time, and Mole Ratio in the Reaction of I with Cyclohexanone.-These results are given in Table I. The following procedure (expt 5) provided 42-45% of the dioxaphospholane IV consistently. A mixture of I (21.0 g, 0.1 mole) and cyclohexanone (31 ml, 0.3 mole) was kept for 3 hr at reflux temperature (internal temperature 155-159'). The unreacted ketone was removed below 60' (18 mm, bath a t 90') and the residue was separated into three fractions by means of a 6-in. Vigreux column: ( 1 ) 4.6 g, bp 35-50' (0.04 mm); (2) 4.11 g, bp 50-86" (0.04 mm); (3) 13.0 g, 42% yield of IV, bp 86-92' (0.04mm), n% 1.4641. Behavior of the Trimethyl Biacetylphosphite 1:1 Adduct (I j at Elevated Temperatures. A.-Pure I was kept 72 hr at 125'. The Pal nmr spectrum of the resulting liquid had the signals due to: (1) recovered 1 : l adduct I , (2) meso- and rac-trimethylbiacetylphosphite 2 : 1 adduct, 2,2,2-trimethoxy-4,5-diacetyl4,5dimethyl-1,3,2-dioxaphospholane,Va and 1%; (3) trimethyl phosphate, (CHa0)aPO; (4) trimethyl phosphite, (CH30)3P. The H1 nmr spectrum confirmed the presence of these substances in the pyrolysis mixture. B.-Sufficient meso and racemic 2 : l adducts (Va and Vb) had been produced after 6 hr at 160°, for detection in the H1 nmr spectrum of the pyrolysis mixture. Behavior of the Trimethyl Biacetylcyclohexanonephosphite Adduct (IV)at Elevated Temperature.-The dioxaphospholane IV underwent significant decomposition after 6 hr a t 165'. The formation of some cyclohexanone was verified, but the mixture was not investigated further. Reaction of the Trimethyl Biacetylphosphite 1:1 Adduct (I) with Cyclobutanone. Preparation of 2,2,2-Trimethoxy-4-acetyl4-methyl-5-spirotrimethylene-l,3,2-dioxaphospholane (IV).-A mixture of 1:1 adduct I (9.4 g) and cyclobutanone (9.4 g, 3 mole equiv) was kept for 40 hr a t reflux temperature. The internal temperature ranged from 99 to 105'. The excess cyclobutanone was recovered a t 96-105' (760 mm) and the rasidue was distilled using a 6-in. Vigreux column: (1) about 50y0 of the 1: 1 adduct I waa recovered a t 30-45' (0.05 mm); (2) 1.l g of dioxaphospholane VI, n 2 b 1.4474, was collected a t 60-62' (0.05
3163
mm); (3) the main portion of VI, 2.1 g, n25~1.4486, was collected at 62-65' (0.05 mm). The combined yield of VI was 40%. Anal. Calcd for ClIHz106P: C, 47.2; H, 7.5; mol wt, 280. Found: C, 46.5; H, 7.3; mol wt, 256 (cryoscopic in benzene). The Pal (CHZCL) and the H1 (cc14) nmr shifts are given in Table 11. The infrared spectrum (CCl4) had a C-0 band at 5.84 and the strong band a t 9.3 p due to the POCHI group. Reaction of the 1:1 Adduct (I) with Hexafluoroacetone. Preparation of 2,2,2-Trimethoxy-4-acetyl-4-methyl-5,5-bistrifluoromethyl-1,3,2-dioxaphospholane(VII).-The gaseous hexafluoroacetone was introduced into a solution of 1 : l adduct I (8.4 g) in hexane (80 ml), cooled in a Dry Ice-acetone bath. After 40 min, the mixture was removed from the bath and was allowed to reach 20' (3 hr). The solvent was removed at 20°, first at 20, then at 1 mm. The residue (14.9 g; theoretical yield, 15.0 g) was distilled through a 6-in. Vigreux column: (1) 0.8 g collected at 34-59' (0.1 mm; bath at 85"); (2) 4.47 g of dioxaphos~ bp 60-62' (0.05 mm; bath at 100'). pholane VII, n Z 51.3911, The yield of VI1 was 90%. Anal. Calcd for C&1.$606P: c , 31.9; H, 4.0; P, 8.2. Found: (2,333; H,4.1; P, 8.5. The Pal (in CH&12) and the H1 (in cc14) nmr shifts are given in Table 11. The infrared spectrum (in CCl4) had a C=O band a t 5.78 and a broad and strong band a t 9.3 p due to the POCHs group. There was a very strong band a t 8.20 p. Reaction of the 1:l Adduct (I) with Acetophenone. Preparation of 2,2,2-Trimethoxy-4a-acetyl-4~,5cu-dimethyl-5p-phenyl1,3,2-dioxaphospholane (VIII).-A mixture of 1 :1 adduct I (21.0 g, 0.1 mole) and acetophenone (36.0 g, 0.3 mole) was kept for a total of 72 hr at 125". The H1 nmr spectra of aliquots were examined after 24 and 48 hr in ccl4. The signal of the acetyl group of the product VI11 was seen to appear at T 7.70 next to the acetyl of acetophenone at 7.52. The methyl groups of the product VI11 appeared at T 9.15 and at 8.55. The excess of monoketone and unreacted I were collected a t 38-44' (0.05 mm), using an 8411. Vigreux column. The dioxaphospholane VI11 (17 g, 50%) was collected a t 90-105" (0.07 mm; bath a t 140-150'). Infrared and H1 nmr spectra showed contamination with small amounts of acetophenone and of 1 :1 adduct I. Pure VIII, bp 110" (0.07 mm; bath at 159'), n% 1.4936, was obtained in ea. 4Oy0 yield after an additional distillation through a l-in. column. Anal. Calcd for C1&&P: C, 54.5; H, 7.0; P, 9.4. Found: C, 54.8; H , 7.2; P,9.4. The P31 (CH2ClZ)and the HI (cc14) nmr shifts are listed in Table 11. The infrared (CC14)spectrum showed a C=O band at 5.84 and the POCHI absorption as a strong, broad, and split band a t 9.25 and 9.32 p . Reaction of the 1: 1 Adduct (I) with p-Nitroacetophenone. Preparation of 2,2,2-Trimethoxy-4a-acetyl-4&5a-dimethyl-5/3p-nitrophenyl-l,3,2-dioxaphospholane(IXa) and of the Diastereomer, 5/3-Methyl-5~-p-nitrophenyl-1,3,2-dioxaphospholane (IXb).-A mixture of I (42.0 g, 200 mmoles) and p-nitroacetophenone (16.5 g, 100 mmoles) was kept for 64 hr at 100". The H1 nmr spectra of aliquots taken after 40 and 60 hr showed little change. The signals due to the diastereomeric phospholanes, IXa and IXb, with trans- and cis-CHs.CO/C&I4NOn-p, respectively, are listed in Table 11. The proportion of isomer was 35:65, translcis. About 5Yc of unreacted p nitroacetophenone was present also. The above mixture was heated to 100' (0.1 mm) to remove the excess phospholene (I) which boiled in the range 40 to 60". The residue was kept for 20 hr at -10 under 30 ml of hexane. The crystalline mixture of isomers IXa and IXb (24 g, 65%) was collected by filtration. The H' nmr spectrum showed the transleis isomers in the proportion of ca. 20: 80 (the trans isomer is more soluble in hexane). The hexane filtrate (12 g after evaporation) contained the balance of IXa IXb. The H1 nmr showed also the presence of some meso- and rac-trimethyl biacetylphosphite 2: 1 adducts (meso, T 7.80 and 8.68; racemic, r 7.70 and 8.76). Some phospholene I was present ( T 8.23). The crystalline mixture of IXa IXb (23 g) was recrystallized from 20 ml of benzene and 30 ml of hexane (16 hr a t 10'). The major cis isomer IXb, mp 91-92', was isolated in 41y0 yield (14 g). Anal. Calcd for C15H22N08P: C, 48.0; H, 5.9; N , 3.7; P,8.3. Found: C,48.3;H,6.1; N, 3.9; P,8.3. The nmr data is given in Table 11. The infrared spectrum (CCla) had strong bands at 5.85 (CO), 6.58 and 7.45, and 9.32 p (phosphorane POCHa).
+
+
VOL. 31
STARNES
3164
The benzene-hexane filtrate, from which 41% of cis isomer (IXb) had been removed, was evaporated. The residue (9 g) consisted of I X a JXb in about 60:40 proportion, with some p-nitroacetophenone as impurity. Pyrolysis of the Trimethyl Biacetyl-p-nitrophenylphosphite Adduct (Mb), cis-CHaCO/C&. NOrp.-The phospholane IXb (2.0 g) was kept for 16 hr at 120', under Nt. The resulting brown mixture was dissolved in CCl4. The H1nmr spectrum of this solution showed the three signals of some unreacted adduct IXb at T 8.37, 8.50, and 8.30. The signals of a small amount of the trans isomer (IXa) were present a t T 7.60, 9.12, and 8.53. There were considerable amounts of p-nitroacetophenone recognized by the acetyl signal at T 7.33. The corresponding dissociation product, the biacetyltrimethylphosphite 1:1 adduct (I), was recognized by its singlet at T 8.20 (CHsC=C) and by its doublet a t 6.50 (CHsO), JHP= 13.0 cps. There were relatively large amounts of trimethylphosphate (doublet at T 6.26, JHP= 11.0 cps). A set of signals a t T 7.82, 8.10, and 7.78 could be due to an epoxide.
+
Reaction of the 1 : 1 Adduct (I) with a,a,a-Trifluoroacetophenone. Preparation of 2,Z,Z-Trimethoxy-4a-acety1-4~-methyl5a-trifluoromethyl-5,6-phenyl-l,3,2-dioxapholane(Xa) and of the Diastereomer, 5,6-Trifluoromethyl-5a-phenyl-1,3,2-dioxaphospholane (Xb).-A mixture of 1 : l adduct I (21.0 g, 100 mmoles) and a,a,a-trifluoroacetophenone (17.4 g, 100 mmoles) was kept for 20 hr at 95'. The course of the reaction was followed by H1 nmr spectroscopy. The mixture was distilled through a 3-in. Vigreux column: (1) forerun, bp 35 to 116' (0.1 mm), 2.5 g; (2) main portion of the two diastereomers Xa Xb, bp 116-118' (0.15 mm, bath a t 165'), 12% 1.4664, yield Xb a t 0.1 mm (bath 25 g, 85%. A second distillation of X a a t 150-155') afforded little separation of stereoisomers; the mixturehad bp 110-112' (0.1 mm), n% 1.4673. Anal. Calcd for CisHzoFaOsP: c, 46.9; H, 5.2; F, 14.8; P,8.1. Found: C,47.9; H,5.4; F, 15.2; P , 8.5. The H1nmr spectrum (in C c 4 ) gave the signals listed in Table 11; integration showed a proportion of 82: 18 for the trans: cis-CHa.CO/CEH6 isomem. The Pal nmr signals (in CH2C12) are listed in Table 11. The infrared spectrum (in CC4) had a C-0 band a t 5.84 and the strong, split band a t 9.22 and 9.40 ~c due to the POCHa group. A strong characteristic double band was present at 8.40 and 8.60 I.(.
+
+
Acid-Catalyzed Decomposition of Peroxydienones Derived from Hindered Phenols' W. H. STARNES, JR. Esso Research and Engineering Company, B a y t m Research and Development Division, Baytown, Texas Received March 88, 1966 Decompositions of 4t-butylperoxy-4methyl-2,6-di-t-butyl-2,5-cyclohexadien-l-one ( la) and 4t-butylperoxy2,4,6-tri-t-butyl-2,5-cyclohexadien-l-one ( l b ) by a variety of Lewis and BrZnsted acids have been carried out at ambient temperature in dilute solution. With sulfuric acid in acetic anhydride l a gives 3,5-di-t-butyl-4acetoxybenzyl acetate (2)in excellent yield, whereas the major product resulting from the reaction of la with boron trifluoride etherate in benzene is 3,5-di-t-butyl-4-hydroxybenzaldehyde (7). Evidence is presented for the intermediacy of quinone methide 3 in these reactions. In contrast, decomposition of l a with hydrogen chloride in benzene gives mostly 2-tbutyl-6-chloro-p-cresol(20). Approximately 90% of 20 is formed by a mechanism that does not require the presence of a chlorinating agent which is able to attack aromatic nuclei. The remaining 10% of 20 results from chlorodealkylation of 2,6-di-t-butyl-pcresol ( 10) by a chlorinating species generated in situ from hydrogen chloride and t-butyl hydroperoxide. Hydrogen chloride catalyzed conversion of chlorodienone 22 to 20 occurs in high yield and proceeds by a mechanism which is entirely different from that involved in the hydrogen halide catalyzed conversion of bromodienone 23 to 2-t-butyl-6-bromo-pcresol (24). Decomposition of l b gave high yields of 2,6-di-t-butylquinone (8) under all of the conditions studied.
It is now generally agreed that inhibition of the lowtemperature liquid phase autoxidation of hydrocarbons by hindered phenols2 involves the formation of peroxycyclohexadienones. Mechanistic studies of inhibited autoxidation have frequently been carried out on the assumption that these peroxides were stable under the reaction conditions, and there seems to be no reason to doubt the validity of this postulate in most of the cases where it has been put forth. On the other hand, it is clear that there are many systems encountered in practical stabilization work where reactions of peroxycyclohexadienones are likely to 0ccur.I Drastic changes in over-all inhibition mech(1) Presented at the Southeast-Southwest Regional Meeting of the American Chemical Society, Memphis, Tenn., Dee 2, 1965. This paper constitutes part I11 of a series on oxidation inhibitors. Parts I and I1 are (a) N. P. Neureiter and D. E. Bown, Ind. Eno. Chem., Prod. Rea. Develop., 1, 236 (1962); (b) N. P. Neureiter, J . Org. Chem., 48, 3486 (1963). (2) The literature through 1960 relating to this subject has been reviewed by K. U. Ingold, Chem. Reo., 61, 563 (1961). (3) For examples of the preparation of alkylperoxycyclohexadienonesand discussions of their role in inhibition, see (a) T. W. Campbell and G. M. Coppinger, J . A m . Chem. Soc., 74, 1469 (1952); (b) A. F. Bickel, E. C. Kooyman, and C. la Lau, J . Chem. Soc., 3211 (1953); (c) C. E. Boozer, G. 9. Hammond, C. E. Hamilton, and J. N . Sen, J . A m . Chem. Soc., 77, 3233 (1955); (d) E. C. Horswill and K. U. Ingold, Can. J . Chem., 44, 263,269 (1966). (4) See ref l a for a discussion of some systems of thin type.
anisms might easily come about as a consequence of these reactions, and for this reason the chemistry of peroxycyclohexadienones is of potential practical interest. The present paper is concerned with the acid-catalyzed5 decompositions of these peroxides, a topic which heretofore has not been subjected to systematic scrutiny. Two representative compounds, la and b, were selected for study, and the objectives of the work were twofold: first, to determine what products are formed from these substances when they are decom-
FT la, R=CH3 b, R = (CH3),C
posed by common Lewis and Br#nsted acids; and second, to arrive at qualitative descriptions of the (5) The term "acid-catalyned" is used here merely for convenience, since there is no guarantee that all of the acids employed in this work functioned a8 true catalysts ( L e . , were not destroyed during reaction).
